BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20241120T082409Z
LOCATION:HG E 1.2
DTSTART;TZID=Europe/Stockholm:20240604T170000
DTEND;TZID=Europe/Stockholm:20240604T173000
UID:submissions.pasc-conference.org_PASC24_sess114_msa179@linklings.com
SUMMARY:SYCLing Through the Portability Maze: Adventures in Scientific App
 lication Migration with DPEcho and HemeLB
DESCRIPTION:Minisymposium\n\nSalvatore Cielo (Leibniz Supercomputing Centr
 e), Alexander Poeppl (Intel Corporation), Ivan Pribec (Leibniz Supercomput
 ing Centre), and Ioannis Zacharoudiou (University College London)\n\nIn ou
 r presentation we explore the portability challenges and solutions encount
 ered in adapting scientific applications to heterogeneous computing archit
 ectures using SYCL. To this end, we present two application porting effort
 s, HemeLB and DPEcho, spanning different scientific domains and requiring 
 different porting strategies. DPEcho is a general relativistic magnetohydr
 odynamics (GR-MHD) code for compact, magnetized astrophysical objects, suc
 h as black hole accretion and stellar winds, or MHD waves. DPEcho is a SYC
 L + MPI rewriting of a legacy CPU-only Fortran application, written with t
 he goals of source readability and cross-platform portability in mind. At 
 the same time, we achieved a high performance on all tested hardware, and 
 outperformed the original implementation on the same CPU architecture. Hem
 eLB is a large-scale Lattice Boltzmann solver developed for simulating blo
 od flow in sparse geometries such as the human vasculature system. It clev
 erly overlaps communication and computation to achieve good scaling on lar
 ge-scale HPC systems. With a CUDA version already available, we pursued SY
 CL porting using the automatic Data Parallel C++ Compatibility (DPCT) tool
 , and then manually refactored sections of the code to account for differe
 nces between programming models. This resulted in a unified code tree whic
 h can execute on all major GPU architectures.\n\nDomain: Engineering, Comp
 utational Methods and Applied Mathematics\n\nSession Chair: Szilárd Páll (
 KTH Royal Institute of Technology, PDC Center for High Performance Computi
 ng)
END:VEVENT
END:VCALENDAR